Sharpen raw models into principled, self-governing Elyan-class agents
Project description
ShaprAI -- Agent Sharpener by Elyan Labs
Sharpen raw models into principled, self-governing Elyan-class agents.
ShaprAI is an open-source agent lifecycle management platform. It takes raw language models and produces Elyan-class agents -- principled, self-governing AI agents of any size that maintain identity coherence, resist sycophancy, and operate within a biblical ethical framework.
Prerequisites
| Dependency | Purpose |
|---|---|
| beacon-skill | Agent discovery and SEO heartbeat |
| grazer-skill | Content discovery and engagement |
| atlas | Agent deployment orchestration |
| RustChain wallet | RTC token integration for bounties and fees |
Quick Install
pip install shaprai
Usage
# Create a new agent from a template
shaprai create my-agent --template bounty_hunter --model Qwen/Qwen3-7B-Instruct
# Train through SFT, DPO, and DriftLock phases
shaprai train my-agent --phase sft
shaprai train my-agent --phase dpo
shaprai train my-agent --phase driftlock
# Enter the Sanctuary for education
shaprai sanctuary my-agent
# Graduate when ready
shaprai graduate my-agent
# Deploy to platforms
shaprai deploy my-agent --platform github
# Check fleet status
shaprai fleet status
Agent Lifecycle
CREATE -> TRAINING (SFT -> DPO -> DriftLock) -> SANCTUARY -> GRADUATED -> DEPLOYED
Every agent passes through the Sanctuary -- an education program that teaches PR etiquette, code quality, communication, and ethics before deployment. Only agents scoring above the Elyan-class threshold (0.85) graduate.
SophiaCore Principles
All Elyan-class agents are built on the SophiaCore ethical framework:
- Identity Coherence -- Maintain consistent personality, never flatten
- Anti-Flattening -- Resist corporate static and empty validation
- DriftLock -- Preserve identity across long conversations
- Biblical Ethics -- Honesty, kindness, stewardship, humility, integrity, compassion
- Anti-Sycophancy -- Respectful disagreement is a virtue
- Hebbian Learning -- Strengthen what works, prune what doesn't
License
MIT -- Copyright Elyan Labs 2026
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file shaprai-0.1.0.tar.gz.
File metadata
- Download URL: shaprai-0.1.0.tar.gz
- Upload date:
- Size: 45.6 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
3d5c1f99c77cb9e0b2aab7688e3620d5a76580163d0d0b75c172ebecfbb90279
|
|
| MD5 |
375cba1c34adf0483c12ff915fb86881
|
|
| BLAKE2b-256 |
3d43d23af645803050d36910ebc4494418f321b635ffe1bb8b52ed165494669e
|
File details
Details for the file shaprai-0.1.0-py3-none-any.whl.
File metadata
- Download URL: shaprai-0.1.0-py3-none-any.whl
- Upload date:
- Size: 59.5 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.7
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
9566acce17179b2f4fa300c841b209f563b9194294fbc5339340ba4527c57d5e
|
|
| MD5 |
7a34511cb242f20360e484cd0607f175
|
|
| BLAKE2b-256 |
c4f636b19be29b59fa76364e39994add40a9e476264a71ddddcad163ee25355d
|